INDEX

Explanations

instances where individuals are being subjected to something, potentially against their will

oai_token-act-pair · gpt-3.5-turbo

instances of the word "subjected" and related concepts indicating experiences of hardship or treatment

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

enegger

-0.65

flies

-0.65

 Islands

-0.62

 Mush

-0.61

AGE

-0.61

assies

-0.60

 Worldwide

-0.60

 beans

-0.59

 Berger

-0.59

 Fargo

-0.59

POSITIVE LOGITS

aton

0.95

ophile

0.93

etics

0.91

ĪĴ

0.90

hran

0.90

stasy

0.83

ħĭ

0.83

yssey

0.81

urst

0.81

itive

0.80

Activations Density 0.033%

instances where individuals are being subjected to something, potentially against their will

instances of the word "subjected" and related concepts indicating experiences of hardship or treatment

No Comments

No Known Activations